https://orcid.org/0000-0001-6875-8940
Abstract
This study aimed to understand the role of mammalian target of rapamycin (mTOR) in CD8+ cells in the pathogenicity of RA and the changes after treatment with biologic drugs.
Peripheral blood mononuclear cells (PBMCs) were isolated from 17 healthy controls and 86 patients with RA. Phosphorylation of mTOR (p-mTOR) and its clinical relevance were evaluated. The role of mTOR in CD8+ cells was also examined in vitro.
Patients with RA who had a moderate or high disease activity, were biologic-naïve, and were refractory to MTX were enrolled in this study. The p-mTOR levels in CD8+ cells were higher in patients with RA than in healthy controls, and they positively correlated with the disease activity in such patients. However, after one year of treatment with TNF inhibitors, the p-mTOR levels in CD8+ cells were suppressed and showed a positive correlation with the treatment response, which was not observed in the abatacept-treatment group. In vitro stimulation of CD8+ cells with anti-CD3 and anti-CD28 antibodies induced mTOR phosphorylation and increased the production of granzyme B, granulysin, TNF-α and IFN-γ but decreased the production of granzyme K. However, on treatment with TNF inhibitors, p-mTOR levels in CD8+ cells and granzyme B production decreased, while granzyme K production increased. The production of granulysin and IFN-γ was not affected by the TNF inhibitors.
These results suggested that mTOR activation in CD8+ cells may be a novel evaluation marker for RA disease activity and a predictive marker of therapeutic response to TNF inhibitors.
p-mTOR level in CD8+ cells was higher in RA patients than in healthy controls.
p-mTOR in CD8+ cells was positively correlated with disease activity and TNF inhibitors-treatment response.
In vitro, mTOR activation in CD8+ cells induced granzyme B production in TNF-α production-dependent manner.
Introduction
Recent single-cell transcriptomics and mass cytometry analyses of the synovial tissue of RA patients have revealed that there are three clusters of CD8+ cells, granzyme K (GZMK)+, granulysin (GNLY)+, granzyme B (GZMB)+ (cytotoxic lymphocytes), and GZMK+GZMB+ cells, which produce IFN-γ. These data reaffirm the importance of CD8+ cells in RA pathology [1]. However, further clarification is required.
HLA-DR and CD38 are widely recognized as activation markers of CD8+ cells, which have been established in human immunology projects [2]. However, it has not been verified whether the proportion of these markers in CD8+ cells is closely related to RA pathogenesis. In addition, it is necessary to discover new biomarkers related to pathological conditions of RA. In recent years, many studies have focussed on immunometabolism. Metabolic reprogramming is essential for the activation and differentiation of immune cells to meet the enormous energy demands, such as adenosine triphosphate (ATP) production and precursor supply, for nucleic acid and lipid biosynthesis [3, 4]. Metabolic pathways are known to play critical roles in the generation of key products that support cell survival and growth [5]. To date, there are six key metabolic pathways, including glycolysis, oxidative phosphorylation, the pentose phosphate pathway, fatty acid β-oxidation, fatty acid synthesis and amino acid pathways/glutaminolysis. Furthermore, mammalian target of rapamycin (mTOR) plays an important role in promoting anabolism, such as enhanced glycolysis [6].
There are many reports on autoimmune diseases, such as SLE, about the involvement of immunometabolism in lymphocytes. Their results have demonstrated the importance of metabolic reprogramming, such as mTOR activation, aerobic glycolysis, fatty acid synthesis and glutaminolysis, in CD4+ cells [7–10]. In our study, we have reported that mTOR activation of B cells correlated with the disease activity in patients with RA and SLE [11, 12]. In addition, glycolysis has been shown to enhance the pentose phosphate pathway in CD4+ cells in patients with RA [13]. However, the involvement of immunometabolism in CD8+ cells in RA remains unknown.
In this study, we aimed to clarify the pathological role of mTOR in CD8+ cells in RA. Specifically, we assessed the changes in the mTOR levels of CD8+ cells after treatment with biologics. Finally, we examined the functional role of mTOR in CD8+ cells in vitro.
Materials and methods
Patients
Peripheral blood mononuclear cells (PBMCs) were obtained from 86 patients with RA and 17 age-/sex-matched healthy controls at the University of Occupational and Environmental Health, Japan. Table 1 lists the clinical characteristics of the study participants. Patients with RA who had more than moderate disease activity [DAS28-ESR ≥3.2 and/or simple disease activity index (SDAI) >11.0], were refractory to MTX treatment (>3 months) and were biologic-naïve were registered with their clinical history. Of the 86 patients with RA initially enrolled, 62 patients followed up with clinical symptoms after 1 year, and 22 patients followed up with clinical symptoms and flow cytometry analysis after 1 year. Biologics (TNF inhibitors and abatacept) were selected through shared decision-making. The collection of peripheral blood samples from the HDs and patients was approved by the Human Ethics Review Committee of the University of Occupational and Environmental Health, Japan, and each subject provided a signed consent form (H23-005).
Characteristics of patients with RA in Fig. 1
| RA (n = 86) | HD (n = 17) | P-value | |
|---|---|---|---|
| Age | 59.4 (16.1) | 52.4 (15.0) | 0.1003 |
| Gender (ratio of female) | 74.4 | 82.4 | 0.4905 |
| Disease duration (months) | 56 | ||
| Stage (I/II/III/IV) | 36/33/8/9 | ||
| Corticosteroid (%) | 13.9 | ||
| Dose of corticosteroid (mg/day) | 5.7 (3.7) | ||
| Dose of MTX (mg/week) | 13.7 (5.8) | ||
| Other csDMARDs | |||
| Tacrolimus | 7 | ||
| Salazosulfapyridine | 7 | ||
| Leflunomide | 1 | ||
| Tender joints (28) | 9.1 (6.0) | ||
| Swollen joints (28) | 7.2 (5.1) | ||
| DAS28 (CRP) | 4.8 (1.3) | ||
| DAS28 (ESR) | 5.8 (3.1) | ||
| CDAI | 26.1 (12.3) | ||
| SDAI | 28.4 (14.3) | ||
| HAQ | 1.2 (0.8) | ||
| CRP | 2.5 (3.4) | ||
| ESR | 47.0 (31.9) | ||
| RF [titer(IU/ml)] | 170.2 (345.0) | ||
| ACPA[titer(U/ml)] | 299.6 (480.5) | ||
| MMP-3 | 243.8 (337.6) | ||
| IgG | 1493.5 (482.5) |
| RA (n = 86) | HD (n = 17) | P-value | |
|---|---|---|---|
| Age | 59.4 (16.1) | 52.4 (15.0) | 0.1003 |
| Gender (ratio of female) | 74.4 | 82.4 | 0.4905 |
| Disease duration (months) | 56 | ||
| Stage (I/II/III/IV) | 36/33/8/9 | ||
| Corticosteroid (%) | 13.9 | ||
| Dose of corticosteroid (mg/day) | 5.7 (3.7) | ||
| Dose of MTX (mg/week) | 13.7 (5.8) | ||
| Other csDMARDs | |||
| Tacrolimus | 7 | ||
| Salazosulfapyridine | 7 | ||
| Leflunomide | 1 | ||
| Tender joints (28) | 9.1 (6.0) | ||
| Swollen joints (28) | 7.2 (5.1) | ||
| DAS28 (CRP) | 4.8 (1.3) | ||
| DAS28 (ESR) | 5.8 (3.1) | ||
| CDAI | 26.1 (12.3) | ||
| SDAI | 28.4 (14.3) | ||
| HAQ | 1.2 (0.8) | ||
| CRP | 2.5 (3.4) | ||
| ESR | 47.0 (31.9) | ||
| RF [titer(IU/ml)] | 170.2 (345.0) | ||
| ACPA[titer(U/ml)] | 299.6 (480.5) | ||
| MMP-3 | 243.8 (337.6) | ||
| IgG | 1493.5 (482.5) |
CDAI: clinical disease activity index; DAS: disease activity index; SDAI: simplified disease activity index.
Characteristics of patients with RA in Fig. 1
| RA (n = 86) | HD (n = 17) | P-value | |
|---|---|---|---|
| Age | 59.4 (16.1) | 52.4 (15.0) | 0.1003 |
| Gender (ratio of female) | 74.4 | 82.4 | 0.4905 |
| Disease duration (months) | 56 | ||
| Stage (I/II/III/IV) | 36/33/8/9 | ||
| Corticosteroid (%) | 13.9 | ||
| Dose of corticosteroid (mg/day) | 5.7 (3.7) | ||
| Dose of MTX (mg/week) | 13.7 (5.8) | ||
| Other csDMARDs | |||
| Tacrolimus | 7 | ||
| Salazosulfapyridine | 7 | ||
| Leflunomide | 1 | ||
| Tender joints (28) | 9.1 (6.0) | ||
| Swollen joints (28) | 7.2 (5.1) | ||
| DAS28 (CRP) | 4.8 (1.3) | ||
| DAS28 (ESR) | 5.8 (3.1) | ||
| CDAI | 26.1 (12.3) | ||
| SDAI | 28.4 (14.3) | ||
| HAQ | 1.2 (0.8) | ||
| CRP | 2.5 (3.4) | ||
| ESR | 47.0 (31.9) | ||
| RF [titer(IU/ml)] | 170.2 (345.0) | ||
| ACPA[titer(U/ml)] | 299.6 (480.5) | ||
| MMP-3 | 243.8 (337.6) | ||
| IgG | 1493.5 (482.5) |
| RA (n = 86) | HD (n = 17) | P-value | |
|---|---|---|---|
| Age | 59.4 (16.1) | 52.4 (15.0) | 0.1003 |
| Gender (ratio of female) | 74.4 | 82.4 | 0.4905 |
| Disease duration (months) | 56 | ||
| Stage (I/II/III/IV) | 36/33/8/9 | ||
| Corticosteroid (%) | 13.9 | ||
| Dose of corticosteroid (mg/day) | 5.7 (3.7) | ||
| Dose of MTX (mg/week) | 13.7 (5.8) | ||
| Other csDMARDs | |||
| Tacrolimus | 7 | ||
| Salazosulfapyridine | 7 | ||
| Leflunomide | 1 | ||
| Tender joints (28) | 9.1 (6.0) | ||
| Swollen joints (28) | 7.2 (5.1) | ||
| DAS28 (CRP) | 4.8 (1.3) | ||
| DAS28 (ESR) | 5.8 (3.1) | ||
| CDAI | 26.1 (12.3) | ||
| SDAI | 28.4 (14.3) | ||
| HAQ | 1.2 (0.8) | ||
| CRP | 2.5 (3.4) | ||
| ESR | 47.0 (31.9) | ||
| RF [titer(IU/ml)] | 170.2 (345.0) | ||
| ACPA[titer(U/ml)] | 299.6 (480.5) | ||
| MMP-3 | 243.8 (337.6) | ||
| IgG | 1493.5 (482.5) |
CDAI: clinical disease activity index; DAS: disease activity index; SDAI: simplified disease activity index.
Cell isolation and differentiation
PBMCs from healthy subjects and patients with RA were immediately isolated using lymphocyte separation medium (Lympholyte-H; Cedarlane, Burlington, NC, USA). CD8+ T cells were isolated from PBMCs using the MojoSort™ Human CD8 T Cell Isolation Kit (BioLegend, San Diego, CA, USA). According to the manufacturer’s instructions, the purity of the isolated CD8+ T cells was confirmed to be >90%, as determined by flow cytometric analysis. Cells (2 × 105 cells/200 µL) were cultured in 96‐well flat‐bottomed plates with complete Roswell Park Memorial Institute (RPMI) 1640 medium (Wako) supplemented with 10% foetal calf serum and 1% penicillin/streptomycin (Life Technologies). All cells were cultured in a humidified incubator at 37°C and 5% CO2. The cells were activated using plate-bound anti‐human CD3 (2 mg/ml; eBioscience) and anti-CD28 (0.5 µg/ml; eBioscience), with or without 10 nM rapamycin (mTORC1 inhibitor) (Selleck Chemicals, Houston, TX, USA), or 2.5 nM LKB1 inhibitor (Pim1/AAK1 dual inhibitor) (CEM, Japan), 2.5 nM AMP-activated protein kinase (AMPK) inhibitor (Dorsomorphin) (Wako, Japan), 10 µg/ml adalimumab (ADA) (AbbVie, US), 10 µg/ml etanercept (ETN) (Pfizer, UK), 10 µg/ml abatacept (ABT) (Bristol-Myers Squibb), 10 µg/ml certolizumab pegol (CZP) (UCB, Inc.) and 10 µg/ml tocilizumab (TCZ) (Chugai, Japan) for 3 days.
Cytokine production
The levels of IFN-γ and TNF-α production in the culture medium were assessed using BD Cytometric Bead Array Human Flex Set (BD) and analysed using FCAP ArrayTM software (version 3.0.1 BD).
Flow cytometry
The cells (PBMCs or CD8+ T cells) were first washed and then suspended in 100 µL of FACS solution (0.05% globulin as an Fc blocking reagent, 0.5% human albumin and 0.1% NaN3 in PBS) and stained with the following antibodies: V450-conjugated anti-CD3 Abs (#560365), and/or V500-conjugated anti-CD4 Abs (#560768), APC-H7-conjugated anti-CD8 Abs (#561423), BV510-conjugated anti-CD38 Abs (#563251), PerCP-conjugated anti-HLA-DR Abs (#307628), PE-conjugated anti-CD80 Abs (#560925), APC-conjugated anti-CD86Abs (#560956) (all antibody conjugates were from either BD or Biolegend) for 30 min at 4°C. For intracellular staining, PE-conjugated mTOR (pS2448) (#563489), Alexa 488-conjugated anti-Granulysin Abs (#558254), PerCP-Cy5.5-conjugated anti-Granzyme K Abs (#370513) and Alexa 700-conjugated anti-granzyme B antibodies (#372221) were used.
PBMCs were fixed and permeabilized with Perm Buffer III (BD) or Transcription Factor Buffer Set (BD) before intracellular staining. Finally, the stained cells were washed three times with FACS solution and analysed using FACSVerse (BD Bioscience) and FlowJo software (Tomy Digital Biology). Isotype-matched mouse IgG controls (BD Biosciences) were used to evaluate the background. ΔMean fluorescence intensity (MFI) is defined as the difference between the values of the measured molecule (e.g. p-mTOR) and the IgG isotype control antibody.
Statistical analyses
All data were expressed as mean (s.d.) and analysed using Prism software. Differences between groups were examined for statistical significance using paired or unpaired t tests. The Mann–Whitney test was used for data that were not normally distributed. The Pearson correlation coefficient was used to test the relationship between the two variables of interest. The χ2 test was performed using JMP9 software (SAS Institute Japan, Tokyo). A P-value of <0.05 denoted the presence of statistical significance.
Results
mTOR activation of CD8+ cells was enhanced in RA patients, correlating with disease activity
We analysed the peripheral blood of 17 healthy controls and 86 age-/sex-matched patients with RA whose clinical characteristics are mentioned in Table 1. The mean age was 59.4 years, and the majority of participating patients were female (74.4%). The mean duration of RA disease was 56 months. Among these patients with RA, 80.2% were at stage I and II, and 13.9% were taking corticosteroids at an average dose of 5.7 mg/day. The average dose of MTX was 13.7 mg/week. The mean disease activity at baseline was 9.1, 7.2, 4.8, 5.8, 26.1 and 28.4 for tender joints (28), swollen joints (28), DAS28-CRP, DAS28-ESR, clinical disease activity index (CDAI), and SDAI, respectively. These patients showed high disease activity. As described above, biologic-naïve patients with moderate or high disease activity refractory to MTX were enrolled in this study. In addition, mean RF and ACPA were 170.2 IU/ml (75.6% positive) and 299.6 U/ml (69.8% positive), respectively.
First, we assessed p-mTOR expression levels among CD8+ cells in healthy controls and patients with RA. In peripheral blood CD8+ cells, the level of p-mTOR in patients with RA was higher than that in healthy controls (Fig. 1A). Patients with RA who were not treated with glucocorticoids also showed the same tendency (Fig. 1B). There was no significant difference between the patients with RA who were ACPA-negative and ACPA-positive (Fig. 1C). Combined with the clinical background, the p-mTOR level of CD8+ cells positively correlated with disease activity scores, such as CDAI and SDAI (Fig. 1D). RA patients who were not treated with glucocorticoids also showed a similar tendency (Supplementary Fig. S1, available at Rheumatology online).
Enhancement of p-mTOR level in CD8+ cells in patients with RA, correlating with disease activity
Peripheral blood mononuclear cells were obtained from 17 HDs and 86 patients with RA, and CD3+CD8+ cells were gated using flow cytometry. (A) Expression of mTOR phosphorylation (p-mTOR) in CD8+ cells of HD and patients with RA. (B) Expression of mTOR phosphorylation (p-mTOR) in CD8+ cells of HD and patients with RA who were not treated with glucocorticoids. (C) Expression of mTOR phosphorylation (p-mTOR) in CD8+ cells of HDs and patients with RA with ACPA-negative or ACPA-positive cells. (D) Use of Pearson correlation coefficient to test the relationship between p-mTOR among CD8+ cells and clinical background in patients with RA. **P < 0.01. HD: healthy donors; mTOR: mammalian target of rapamycin.
HLA-DR and CD38 were originally reported as activation markers in CD8+ cells [2]. Therefore, we also assessed the proportion of HLA-DR+CD38+ cells among CD8+ cells, which was found to be higher in patients with RA than in healthy controls (Supplementary Fig. S2A, available at Rheumatology online). Patients with RA who were not treated with glucocorticoids also showed the same tendency (Supplementary Fig. S2B, available at Rheumatology online), and the proportion of HLA-DR+CD38+CD8+ cells in them showed a positive correlation with patient age, disease duration, RF, ACPA, and disease activity scores, such as CDAI and SDAI (Supplementary Fig. S2C, available at Rheumatology online). However, in peripheral blood CD8+ cells of patients with RA who were not treated with glucocorticoids, the expression level of p-mTOR was not correlated with the proportion of CD38+HLA-DR+ cells (Supplementary Fig. S2D, available at Rheumatology online).
mTOR activation in CD8+ cells was suppressed by TNF inhibitors, correlating with the treatment response
Next, we examined the effects of biologics [TNF inhibitors and abatacept (ABT)] on the expression levels of p-mTOR in CD8+ cells and the proportion of HLA-DR+CD38+CD8+ cells in patients with RA. The basic characteristics of the patients with RA in Fig. 2 and Supplementary Fig. S3 (available at Rheumatology online) are described in Supplementary Table S1 (available at Rheumatology online). SDAI, which is an index of disease activity, decreased significantly after one year of treatment with TNF inhibitors or ABT (Fig. 2A). On analysing all the cases together, the average expression level of p-mTOR in CD8+ cells was found to be significantly reduced after the treatment. However, when the patients were divided into two groups, the expression level of p-mTOR decreased significantly in the TNF inhibitor-treatment group, but there was no change in the ABT-treatment group (Fig. 2B). Although there was a significant difference only in the age of the patients with RA shown in Fig. 2 (Supplementary Table S1, available at Rheumatology online), the level of p-mTOR expression in CD8+ cells in them was not related to age (Fig. 1D), indicating that age did not affect the results. The ratio of HLA-DR+CD38+ cells among CD8+ cells was significantly reduced after one year in both the ABT-treated and TNF inhibitor-treated groups (Supplementary Fig. S3A, available at Rheumatology online). Analysis of absolute numbers showed a similar tendency (Supplementary Fig. S3B, available at Rheumatology online).
Changes of p-mTOR level among CD8+ cells in patients with RA after treatment with biologics
(A) Clinical disease activity index of patients with RA treated with a total of TNF inhibitors and abatacept (n = 22), TNF inhibitors (n = 12) and abatacept (n = 10) before and one year later. (B) The level of mTOR phosphorylation in CD8+ cells in patients with RA after treatment with TNF inhibitors or abatacept. **P < 0.01, ***P < 0.001. mTOR: mammalian target of rapamycin.
Subsequently, the expression of p-mTOR in CD8+ cells and the effect of treatment with biologics on the expression was investigated. The basic characteristics of the patients with RA in Fig. 3 are described in Supplementary Table S2, available at Rheumatology online. In the TNF inhibitor-treated group, patients with higher p-mTOR expression levels before treatment showed a better response to the therapy, whereas this was not observed in the ABT-treatment group (Fig. 3A). However, there were significant differences in age, sex, disease duration, stage, csDMARD (TAC), p-mTOR expression level in CD8+ cells, and percentage of HLA-DR+CD38+ cells between the TNF inhibitor-treatment group and abatacept-treatment group of RA patients (Fig. 3A, B) (Supplementary Table S2, available at Rheumatology online). Therefore, we analysed the results using two additional methods. First, sensitivity analysis showed a significant negative correlation between p-mTOR levels before treatment and the percentage change in SDAI after 6 months in the TNF inhibitor-treatment group, while there was no correlation in the abatacept-treatment group (Supplementary Table S3, available at Rheumatology online). Further analysis, using propensity score-based inverse probability of treatment weighting (IPTW), minimizing selection bias, and adjusting for patient background, showed similar results (TNF inhibitor-treatment group; r = –0.5404, P = 0.0307, abatacept-treatment group; r = –0.1895, P = 0.6253) (adjusted patient backgrounds are shown in Supplementary Fig. S4, available at Rheumatology online). Furthermore, no correlation was found between the proportion of HLA-DR+CD38+CD8+ cells and the therapeutic response of CD8+ cells to biologics (Fig. 3B). Using a cut-off value of 899.7 (average of Δp-mTOR in patients with RA), the proportion of SDAI remission at 12 months after TNF inhibitor treatment was found to be higher in the CD8+ cell group having high ΔmTOR (Fig. 3C).
Correlation of p-mTOR level among CD8+ cells with TNF inhibitors treatment responsiveness
Correlation of proportions of (A) mTOR phosphorylation level and (B) CD38+HLADR+ cells among CD8+ cells before treatment with ΔSDAI (change from baseline to 6 months after treatment). (C) Use of chi-square test for the relationship between p-mTOR level and SDAI remission rate in the TNF inhibitor-treatment group using JMP9 software (SAS Institute Japan, Tokyo, Japan). mTOR: mammalian target of rapamycin; SDAI: simplified disease activity index.
mTOR activation in CD8+ cells induced granzyme B production in a TNF-α production-dependent manner in vitro
To investigate the functional role of mTOR in CD8+ cells and the direct effect of TNF inhibitors on those cells, CD8+ cells were collected from the peripheral blood of healthy donors and patients with RA and then evaluated after culturing for 72 h. Stimulation of the cells with anti-CD3 and anti-CD28 antibodies induced mTOR phosphorylation, thereby increasing the production of granzyme B, GNLY and IFN-γ and decreasing the production of granzyme K (Fig. 4A and B). This increased production of granzyme B and IFN-γ was suppressed by mTOR inhibitor (rapamycin), but not by other metabolic inhibitors, such as AMPK inhibitor (dorsomorphin) and LKB1 inhibitor (Fig. 4B), suggesting that mTOR activation in CD8+ cells was important for the production of those metabolites. The proportion of p-mTOR+ granzyme B+ cells significantly increased by stimulation with anti-CD3 and anti-CD28 antibodies in CD8+ cells from patients with RA, compared with healthy donors (Fig. 4C).
mTOR activation induced granzyme B and IFN-γ production in human CD8+ cells
CD8+ cells from HD and RA patients were cultured for 3 days with or without plate-bound anti-human CD3 (2 mg/mL) and anti-CD28 (0.5 µg/mL), in addition to the indicated reagents. (A) Double staining of GZMB, GNLY, IFN-γ and GZMK with p-mTOR analysed using flow cytometry. (C) The proportions of GZMB+, GNLY+ and GZMK+ cells using flow cytometry and the concentration of IFN-γ produced by CBA. (A, B) CD8+ cells isolated from the HDs were used. (C) Proportions of p-mTOR+GZMB+ cells in CD8+ cells from HD or patients with RA. Data are represented as the mean (s.d.) of three independent experiments from three different donors. *P < 0.05, **P < 0.01, ***P < 0.001. GNLY: granulysin; GZMB: granzyme B; GZMK: granzyme K; HD: healthy donors; mTOR: mammalian target of rapamycin.
The production of TNF-α induced by anti-CD3 and anti-CD28 antibodies was suppressed by rapamycin, but not by other metabolic inhibitors, such as DM and LKB1 inhibitor (Fig. 5A). mTOR phosphorylation in CD8+ cells was significantly suppressed by TNF inhibitors, certolizumab-pegol and adalimumab, but not by anti-IL-6R antibody (tocilizumab) and cytotoxic T-lymphocyte associated antigen 4 (CTLA4) fusion protein (abatacept) (Fig. 5B). In other words, mTOR activation induced TNF-α production in CD8+ cells, while TNF inhibitors suppressed mTOR activation in both directions. These results suggested positive feedback for mTOR activation and TNF-α production.
mTOR activation in CD8+ cells induced granzyme B production in a TNF-α production-dependent manner
CD8+ cells from HD and patients with RA were cultured for three days with or without plate-bound anti-human CD3 (2 mg/mL) and anti-CD28 (0.5 µg/mL), in addition to the indicated reagents. (A) TNF-α production by CBA. (B) Histograms and ΔMFI of p-mTOR in CD8+ cells. (C) Proportions of GZMB+ cells, GNLY+ GZMK+ cells, and (D) IFN-γ production. (E) Expression of CD80 and CD86 in CD8+ cells. (A-E) CD8+ cells isolated from the HDs were used. (A, B, D) Data are the mean (s.d.) of three independent experiments from three different donors. (C) Data are the mean (s.d.) of four independent experiments from four different donors. (E) Data are the mean (s.d.) of five independent experiments from five different donors. *P < 0.05, **P < 0.01. GNLY: granulysin; GZMB: granzyme B; GZMK: granzyme K; HD: healthy donors; mTOR: mammalian target of rapamycin.
Finally, the effects of various biologics on the production of granzyme B, GNLY, GZMK and IFN-γ were studied in CD8+ cells. Granzyme B production induced by anti-CD3 and anti-CD28 antibodies was suppressed by TNF inhibitors but was not affected by anti-IL-6R antibody and CTLA4 fusion protein. However, granzyme K production was increased by TNF inhibitors. Biologics did not alter the production of GNLY and IFN-γ, although the TNF inhibitor suppressed mTOR activation (Fig. 5C, D). CD8+ cells were stimulated with anti-CD3 and anti-CD28 antibodies, and the expression of CD80 and CD86 was upregulated. CD80 expression was suppressed with certolizumab-pegol and adalimumab treatment, whereas no change was observed on treatment with rapamycin, abatacept or tocilizumab. On the other hand, CD86 expression was inhibited by rapamycin, but not by other conditions (Fig. 5E). Above all, these results suggested that TNF inhibitors might have specifically suppressed granzyme B production through mTOR regulation in CD8+ cells.
Discussion
In patients with RA, the expression of p-mTOR in CD8+ cells increased, showing a positive correlation with disease activity. The mTOR activation level in the pre-treated CD8+ cells was suppressed by TNF inhibitors and showed a positive correlation with the treatment response. In vitro, mTOR activation in CD8+ cells induced granzyme B production in a TNF-α production-dependent manner. These results suggested that mTOR activation in CD8+ cells might be a new marker for the disease activity of RA and a predictive marker of therapeutic response to TNF inhibitors.
Many studies have reported about the abnormalities of cytotoxic CD8+ T cells and their involvement in RA pathology using mice; only a few of them have used human subjects. In patients with RA, the proportion of memory CD8+CD45RO+ cells correlated with the antibody titre of IgM-rheumatoid factor (IgM-RF) [14]. In the synovial tissue of such patients, Tc1 cells produced a large amount of IFN-γ and IL-10 at the same time [15]. Cytokine (TNF-α, IFN-γ and IL-17A) production by CD8+ cells collected from the peripheral blood of RA patients was positively correlated with DAS28. In addition, the study of CD8+ cells in synovial fluid from patients with RA expressed more robust effector memory (CD27+CD62L–) and activated (CD69+) profiles [16]. Although the association between cytokine production, the stage of differentiation, and pathology has been reported, the role of CD8+ cells in the pathology of RA remains unclear.
mTOR plays an important role in promoting anabolism [17]. There are some reports on T-cell metabolism in patients with RA. In the T cells derived from patients with RA, 6-phosphofructo-2-kinase/fructose-2,6-bisphosphase-3 (PFKFB3) was not enhanced upon activation, glycolysis was suppressed and pyruvate/lactate production was reduced [18, 19]. Conversely, glucose-6-phosphate dehydrogenase (G6PD) was enhanced, and there was a shift in the pentose phosphate pathway (PPP). The imbalance in the 6-phosphofructo-1-kinase (PFK)/G6PD ratio led to the overproduction of NADPH and accumulation of reduced glutathione, leading to a reductive intracellular environment [20, 21]. Similarly, activation of mTOR in CD8+ cells may be related to abnormalities in cellular metabolism in RA. The involvement of mTOR in the pathogenesis of RA in the synovial tissues of patients has been reported in a few studies. Rapamycin or rapalog (a rapamycin derivative) suppressed arthritis in animal models [22] and in patients with RA [23] and JIA [24]. mTOR inhibition by sirolimus or everolimus in a TNF-transgenic mouse model improved arthritis and histologically suppressed synovitis [22]. mTOR activity was not significantly altered in CD4+ cells from patients with RA [18, 25], and the effect of mTOR inhibition on arthritis was thought to be mainly due to the inhibition of mTORC1 activation in fibroblast-like synoviocytes (FLS) [26]. However, we have shown that p-mTOR was upregulated in CD8+ cells from patients with RA and correlated with disease activity (Fig. 1A, D). Analysis of the transcriptome of mouse CD8+ cells affected by mTOR inhibition by rapamycin using RNA-Seq has shown that it inhibits several genes related to apoptosis and migration [27]. Comprehensively considering previous reports, mTOR inhibition may affect not only FLS but also CD8+ cells to inhibit joint destruction in RA.
To obtain a better understanding of the differential effects of TNF inhibitors and abatacept on CD8+ cells in RA patients and their involvement in the pathogenesis, we examined the differential effects of these biologics on mTOR phosphorylation and HLA-DR+CD38+ cells as follows. First, both mTOR phosphorylation level and percentage of HLA-DR+CD38+ cells in CD8+ cells were significantly higher in patients with RA than in healthy subjects (Fig. 1A, B, Supplementary Fig. S2A, available at Rheumatology online). The percentage of HLA-DR+CD38+CD8+ cells in CD8+ cells was significantly correlated with age, disease duration, disease activity (CDAI, SDAI) and autoantibodies (RF, ACPA) (Supplementary Fig. S2C, available at Rheumatology online). The percentage of HLA-DR+CD38+ in CD8+ cells was reduced by both TNF inhibitors and abatacept (Supplementary Fig. S3B, available at Rheumatology online), whereas mTOR phosphorylation was reduced only by TNF inhibitors (Fig. 2B). There was no correlation between the level of mTOR phosphorylation and the percentage of HLA-DR+CD38+ cells in CD8+ cells (Supplementary Fig. 2D, available at Rheumatology online). It has been reported that cross-linking of CD38 with specific monoclonal antibodies in human T cells in vitro triggered various signal transductions and induced T-cell activation, proliferation and cytokine secretion [28–30]. Based on these results, CD38 has long been considered an activation marker of T cells. However, CD38 is not a simple activation marker. For example, it was recently reported that the frequency of HLA-DR+CD38+ in CD8+ cells transiently increased during viral clearance in the peripheral blood of COVID-19 patients and positively correlated with improved patient outcomes [31]. On the other hand, CD38 may reduce the anti-tumour effect by inducing metabolic abnormalities [31, 32] and as a marker of terminally exhausted CD8+ cells that are resistant to functional recovery by PD-1 inhibition [33, 34]. In other words, the function of CD38+CD8+ cells appears to be very complicated and diverse. Presumably, HLA-DR+CD38+-activated CD8+ cells and mTOR-phosphorylated CD8+ cells are distinct cell populations that play important roles in RA pathogenesis, but they are functionally overlapping and partly independent. In other words, mTOR-phosphorylated CD8+ cells may be particularly involved in TNF-α-dependent RA pathology.
Recently, three clusters, GZMK+, GNLY+GZMB+ (cytotoxic lymphocytes) and GZMK+GZMB+ cells, which produce IFN-γ were identified using single-cell transcriptomics and mass cytometry analysis of cytotoxic CD8+ cells in the synovial tissues of patients with RA [1]. CD8+ cells are important functional effectors because their cytotoxic function is mediated by (i) MHC class I (granzyme production) and (ii) cytokine production (IFN-γ, TNF-α, IL-17A and IL-10). However, the role of CD8+ cells in RA pathology and the pathological significance of their major effector functions are unknown.
The involvement of granzyme B in RA pathogenesis has been previously reported as follows. In vitro studies have shown that granzyme B has enzymatic activity to cleave aggrecan proteoglycans in cultured cartilage matrix and whole cartilage explants [35, 36]. Many granzyme B-positive cells are present in the chondrocytes of pannus lesions [36]. Furthermore, the level of granzyme B in the synovial fluid of patients with RA is higher than that in healthy subjects [37] and is associated with an early development of erosion, as sighted on X-rays [38]. Genetic mutations in GZMB (rs854350) are also associated with susceptibility to RA [39]. In an animal model of collagen-induced RA in rats, GZMB gene suppression regulated the expression of factors associated with inflammation, apoptosis, and angiogenesis, and alleviated synovial tissue hyperplasia and cartilage tissue damage [40]. Thus, it has been suggested that granzyme B may be deeply involved in the pathogenesis of RA through joint destruction and other mechanisms. In our in vitro study, mTOR activation in CD8+ cells was unrelated to GZMK production, but was strongly involved in GZMB production (Fig. 4A, B). In recent years, GZMK+CD8+ cells have been shown to be more age-related than GZMB+CD8+ cells [41]. However, functional differences between GZMB and GZMK in RA need to be further studied.
A limitation of this study was that although mTOR activation in CD8+ cells correlated with disease activity and therapeutic responsiveness of TNF inhibitors, it did not show a direct causal relationship with the disease. It has been reported that the regulation of CD8 memory T cells using CD28 inhibition and mTOR inhibitors was suppressed after organ transplantation [42, 43]. Additionally, only part of the CD8+ cells and mTOR activation experiments were performed using patient specimens in this study. In the future, it will be necessary to show a direct causal relationship with the disease by using an RA mouse model that selectively lacks mTOR in CD8+ cells. In this study, we also showed that TNF inhibitors might have affected granzyme production by acting directly on CD8+ cells and affecting disease activity. However, in actual in vivo situations, TNF inhibitors may indirectly affect CD8+ cells through their effects on other cells in the synovium. Thus, it is necessary to verify this using the RA mouse model in the future.
These results suggest that mTOR activation in CD8+ cells might be useful as a new disease marker for RA disease activity and a predictive marker of therapeutic response to TNF inhibitors.
Acknowledgements
The authors thank the patients and healthy volunteers for their cooperation and consent to participate in the study. The authors also thank Ms N. Sakaguchi for providing excellent technical assistance.
M.Z. and S.I. wrote the manuscript and designed the experiments. M.Z., S.I., K.So., M.U., Y.F., J.A., Y.M., N.O., M.H.S., Y.T., H.M., A.N., R.K., H.H., G.T., S.L., S.N. and K.Sa. performed the experiments. M.Z., S.I. and K.So. analysed the data. Y.T. supervised the project.
Funding: This work was supported in part by JSPS KAKENHI grant number #JP16K09928 and the University of Occupational and Environmental Health, Japan, through UOEH Grant for Advanced Research (#H29-903 and #H30-905).
Disclosure statement: J.A. and K.Sa. are employees of Mitsubishi Tanabe Pharma. S.N. has received speaking fees from Bristol Myers, Sanofi, Abbvie, Eisai, Eli Lilly, Chugai, Pfizer, Takeda, as well as research grants from Mitsubishi Tanabe, Novartis and MSD. Y.T. received research grants from Mitsubishi-Tanabe, Takeda, Daiichi-Sankyo, Chugai, Bristol-Myers, MSD, Astellas, Abbvie, and Eisai. All other authors have declared no conflict of interest.
Study approval: The collection of peripheral blood samples from the HDs and patients was approved by the Human Ethics Review Committee of the University of Occupational and Environmental Health, Japan, and all patients gave written informed consent (H23-005).
Data availability statement
All data generated or analysed during this study are included in this article (and its supplementary information files).
Supplementary data
Supplementary data are available at Rheumatology online.
Comments